Re: Random generator

--Apple-Mail-3--117896169
Content-Transfer-Encoding: 7bit
Content-Type: text/plain;
	charset=US-ASCII;
	delsp=yes;
	format=flowed

Here's the patch in diff -u format. Also, the comment about the  
algorithm has been corrected in ccl/l0-number


--- ccl/level-0/l0-numbers.lisp.~1.31.~	2007-04-09 08:39:15.000000000  
+0200
+++ ccl/level-0/l0-numbers.lisp	2007-09-01 12:14:56.000000000 +0200
@@ -1770,7 +1770,7 @@
      ab mod m = u + v                   :  u+v < 2^n
      ab mod m = ((u + v) mod 2^n) + 1   :  u+v >= 2^n
-What we do is use 2b and 2n so we can do arithemetic mod 2^32  
instead of
+What we do is use 2b and 2^n so we can do arithemetic mod 2^32  
instead of
2^31.  This reduces the whole generator to 5 instructions on the  
680x0 or
80x86, and 8 on the 60x.
@@ -1782,7 +1782,7 @@
(defun %next-random-pair (high low)
    (let* ((n (nth-value
               1
-             (%multiply 42871 (dpb (ldb (byte 15 0) high)
+             (%multiply 48271 (dpb (ldb (byte 15 0) high)
                                        (byte 16 16)
                                        (ldb (byte 16 0) low ))))))
      (values (ldb (byte 16 16) n)



--- ccl/level-0/PPC/ppc-numbers.lisp.~1.16.~	2007-04-09  
08:39:15.000000000 +0200
+++ ccl/level-0/PPC/ppc-numbers.lisp	2007-09-01 12:22:19.000000000 +0200
@@ -312,7 +312,7 @@
      ab mod m = u + v                   :  u+v < 2^n
      ab mod m = ((u + v) mod 2^n) + 1   :  u+v >= 2^n
-What we do is use 2b and 2n so we can do arithemetic mod 2^32  
instead of
+What we do is use 2b and 2^n so we can do arithemetic mod 2^32  
instead of
2^31.  This reduces the whole generator to 5 instructions on the  
680x0 or
80x86, and 8 on the 60x.



--Apple-Mail-3--117896169
Content-Transfer-Encoding: 7bit
Content-Type: application/octet-stream; x-unix-mode=0644; name=l0-numbers.patch
Content-Disposition: attachment;
	filename=l0-numbers.patch

--- ccl/level-0/l0-numbers.lisp.~1.31.~	2007-04-09 08:39:15.000000000 +0200
+++ ccl/level-0/l0-numbers.lisp	2007-09-01 12:14:56.000000000 +0200
@@ -1770,7 +1770,7 @@
     ab mod m = u + v                   :  u+v < 2^n
     ab mod m = ((u + v) mod 2^n) + 1   :  u+v >= 2^n
 
-What we do is use 2b and 2n so we can do arithemetic mod 2^32 instead of
+What we do is use 2b and 2^n so we can do arithemetic mod 2^32 instead of
 2^31.  This reduces the whole generator to 5 instructions on the 680x0 or
 80x86, and 8 on the 60x.
 
@@ -1782,7 +1782,7 @@
 (defun %next-random-pair (high low)
   (let* ((n (nth-value
              1
-             (%multiply 42871 (dpb (ldb (byte 15 0) high)
+             (%multiply 48271 (dpb (ldb (byte 15 0) high)
                                       (byte 16 16)
                                       (ldb (byte 16 0) low ))))))
     (values (ldb (byte 16 16) n)
--- ccl/level-0/PPC/ppc-numbers.lisp.~1.16.~	2007-04-09 08:39:15.000000000 +0200
+++ ccl/level-0/PPC/ppc-numbers.lisp	2007-09-01 12:22:19.000000000 +0200
@@ -312,7 +312,7 @@
     ab mod m = u + v                   :  u+v < 2^n
     ab mod m = ((u + v) mod 2^n) + 1   :  u+v >= 2^n
 
-What we do is use 2b and 2n so we can do arithemetic mod 2^32 instead of
+What we do is use 2b and 2^n so we can do arithemetic mod 2^32 instead of
 2^31.  This reduces the whole generator to 5 instructions on the 680x0 or
 80x86, and 8 on the 60x.
 

--Apple-Mail-3--117896169
Content-Transfer-Encoding: 7bit
Content-Type: text/plain;
	charset=US-ASCII;
	format=flowed



-- 
Bernd




--Apple-Mail-3--117896169
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Bug-openmcl mailing list
[email protected]
http://clozure.com/mailman/listinfo/bug-openmcl

--Apple-Mail-3--117896169--